Re: [PATCH RFC v4 01/14] [media] Add common video interfaces OF bindings documentation
Hi Laurent,
On 01/25/2013 02:52 AM, Laurent Pinchart wrote:
+Data interfaces on all video devices are described by their child 'port'
+nodes. Configuration of a port depends on other devices participating in
+the data transfer and is described by 'endpoint' subnodes.
+
+dev {
+ #address-cells = 1;
+ #size-cells = 0;
+ port@0 {
+ endpoint@0 { ... };
+ endpoint@1 { ... };
+ };
+ port@1 { ... };
+};
+
+If a port can be configured to work with more than one other device on
+the same bus, an 'endpoint' child node must be provided for each of
+them. If more than one port is present in a device node or there is more
+than one endpoint at a port, a common scheme, using '#address-cells',
+'#size-cells' and 'reg' properties is used.
Wouldn't this cause problems if the device has both video ports and a
child bus ? Using #address-cells and #size-cells for the video ports would
prevent the child bus from being handled in the usual way.
Indeed, it looks like a serious issue in these bindings.
A possible solution would be to number ports with a dash instead of a @,
as done in pinctrl for instance. We would then get
port-0 {
endpoint-0 { ... };
endpoint-1 { ... };
};
port-1 { ... };
One problem here is that index of the port or the endpoint node can have
random value and don't need to start with 0, which is the case for the pinctrl
properties. It makes iterating over those nodes more difficult, instead
of using standard functions like of_node_cmp() we would need to search for
sub-strings in the node name.
Sounds like a good alternative, I can't think of any better solution at the
moment.
+Two 'endpoint' nodes are linked with each other through their
+'remote-endpoint' phandles. An endpoint subnode of a device contains
+all properties needed for configuration of this device for data exchange
+with the other device. In most cases properties at the peer 'endpoint'
+nodes will be identical, however they might need to be different when
+there is any signal modifications on the bus between two devices, e.g.
+there are logic signal inverters on the lines.
+
+Required properties
+---
+
+If there is more than one 'port' or more than one 'endpoint' node
following +properties are required in relevant parent node:
+
+- #address-cells : number of cells required to define port number,
should be 1.
+- #size-cells: should be zero.
I wonder if we should specify whether a port is a data sink or data
source. A source can be connected to multiple sinks at the same time, but
a sink can only be connected to a single source. If we want to perform
automatic sanity checks in the core knowing the direction might help.
Multiple sources can be linked to a single sink, but only one link can be
active at any time.
So I'm not sure if knowing if a DT port is a data source or data sink would
let us to validate device tree structure statically in general.
Such source/sink property could be useful later at runtime, when data
pipeline is set up for streaming.
Yes, I was mostly thinking about runtime.
How do you think this could be represented ? By just having boolean
properties like: 'source' and 'sink' in the port nodes ? Or perhaps in the
endpoint nodes, since some devices might be bidirectional ? I don't recall
any at the moment though.
Source and sink properties would do. We could also use a direction property
that could take sink, source and bidirectional values, but that might be more
complex.
Since we're going to allow multiple endpoints at a port to be active at any
time, for the reasons we discussed in IRC [1], I assume it's no longer
possible to perform sanity checks mentioned above in the core. Should we
then keep the 'source', 'sink' properties in the port nodes ?
[1] http://linuxtv.org/irc/v4l/index.php?date=2013-01-29
I don't think we will have bidirectional link (as that would most probably
involve a very different kind of bus, and thus new bindings).
--
Thanks,
Sylwester
--
To unsubscribe from this list: send the line unsubscribe linux-samsung-soc in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH RFC v4 01/14] [media] Add common video interfaces OF bindings documentation
Hi,
On 01/30/2013 01:40 PM, Sylwester Nawrocki wrote:
On 01/25/2013 02:52 AM, Laurent Pinchart wrote:
+Data interfaces on all video devices are described by their child 'port'
+nodes. Configuration of a port depends on other devices participating in
+the data transfer and is described by 'endpoint' subnodes.
+
+dev {
+ #address-cells =1;
+ #size-cells =0;
+ port@0 {
+ endpoint@0 { ... };
+ endpoint@1 { ... };
+ };
+ port@1 { ... };
+};
+
+If a port can be configured to work with more than one other device on
+the same bus, an 'endpoint' child node must be provided for each of
+them. If more than one port is present in a device node or there is more
+than one endpoint at a port, a common scheme, using '#address-cells',
+'#size-cells' and 'reg' properties is used.
Wouldn't this cause problems if the device has both video ports and a
child bus ? Using #address-cells and #size-cells for the video ports would
prevent the child bus from being handled in the usual way.
Indeed, it looks like a serious issue in these bindings.
A possible solution would be to number ports with a dash instead of a @,
as done in pinctrl for instance. We would then get
port-0 {
endpoint-0 { ... };
endpoint-1 { ... };
};
port-1 { ... };
Another possible solution could be putting the port nodes under
a ports node, for these cases where a device has a child bus.
Where there is no conflict, the ports node could be omitted for
simplicity. Does it sound reasonable ?
device {
ports {
#address-cells = 1;
#size-cells = 0;
port@0 {
endpoint@0 {
reg = 0;
};
endpoint@1 { ... };
};
port@1 { ... };
};
};
One problem here is that index of the port or the endpoint node can have
random value and don't need to start with 0, which is the case for the pinctrl
properties. It makes iterating over those nodes more difficult, instead
of using standard functions like of_node_cmp() we would need to search for
sub-strings in the node name.
--
Thanks,
Sylwester
--
To unsubscribe from this list: send the line unsubscribe linux-samsung-soc in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH RFC v4 01/14] [media] Add common video interfaces OF bindings documentation
Hi Laurent,
Thanks for the review.
On 01/24/2013 11:16 AM, Laurent Pinchart wrote:
[...]
+Data interfaces on all video devices are described by their child 'port'
+nodes. Configuration of a port depends on other devices participating in
+the data transfer and is described by 'endpoint' subnodes.
+
+dev {
+#address-cells = 1;
+#size-cells = 0;
+port@0 {
+endpoint@0 { ... };
+endpoint@1 { ... };
+};
+port@1 { ... };
+};
+
+If a port can be configured to work with more than one other device on the
+same bus, an 'endpoint' child node must be provided for each of them. If
+more than one port is present in a device node or there is more than one
+endpoint at a port, a common scheme, using '#address-cells', '#size-cells'
+and 'reg' properties is used.
Wouldn't this cause problems if the device has both video ports and a child
bus ? Using #address-cells and #size-cells for the video ports would prevent
the child bus from being handled in the usual way.
Indeed, it looks like a serious issue in these bindings.
A possible solution would be to number ports with a dash instead of a @, as
done in pinctrl for instance. We would then get
port-0 {
endpoint-0 { ... };
endpoint-1 { ... };
};
port-1 { ... };
Sounds like a good alternative, I can't think of any better solution at the
moment.
+Two 'endpoint' nodes are linked with each other through their
+'remote-endpoint' phandles. An endpoint subnode of a device contains all
+properties needed for configuration of this device for data exchange with
+the other device. In most cases properties at the peer 'endpoint' nodes
+will be identical, however they might need to be different when there is
+any signal modifications on the bus between two devices, e.g. there are
+logic signal inverters on the lines.
+
+Required properties
+---
+
+If there is more than one 'port' or more than one 'endpoint' node following
+properties are required in relevant parent node:
+
+- #address-cells : number of cells required to define port number, should
be 1.
+- #size-cells: should be zero.
I wonder if we should specify whether a port is a data sink or data source. A
source can be connected to multiple sinks at the same time, but a sink can
only be connected to a single source. If we want to perform automatic sanity
checks in the core knowing the direction might help.
Multiple sources can be linked to a single sink, but only one link can be
active at any time.
So I'm not sure if knowing if a DT port is a data source or data sink would
let us to validate device tree structure statically in general.
Such source/sink property could be useful later at runtime, when data pipeline
is set up for streaming.
How do you think this could be represented ? By just having boolean
properties like: 'source' and 'sink' in the port nodes ? Or perhaps in the
endpoint nodes, since some devices might be bidirectional ? I don't recall
any at the moment though.
+Optional endpoint properties
+
+
+- remote-endpoint: phandle to an 'endpoint' subnode of the other device
+ node.
+- slave-mode: a boolean property, run the link in slave mode.
+ Default is master mode.
What are master and slave modes ? It might be worth it describing them.
This was originally proposed by Guennadi, I think he knows exactly what's
the meaning of this property. I'll dig into relevant documentation to
find out and provide more detailed description.
+- bus-width: number of data lines, valid for parallel busses.
+- data-shift: on parallel data busses, if bus-width is used to specify the
+ number of data lines, data-shift can be used to specify which data lines
+ are used, e.g. bus-width=10; data-shift=2; means, that lines 9:2
+ are used.
+- hsync-active: active state of HSYNC signal, 0/1 for LOW/HIGH
+ respectively.
+- vsync-active: active state of VSYNC signal, 0/1 for LOW/HIGH
+ respectively. Note, that if HSYNC and VSYNC polarities are not
+ specified, embedded synchronization may be required, where supported.
+- data-active: similar to HSYNC and VSYNC, specifies data line polarity.
+- field-even-active: field signal level during the even field data
+ transmission.
+- pclk-sample: sample data on rising (1) or falling (0) edge of the pixel
+ clock signal.
+- data-lanes: an array of physical data lane indexes. Position of an entry
+ determines the logical lane number, while the value of an entry indicates
+ physical lane, e.g. for 2-lane MIPI CSI-2 bus we could have
+ data-lanes = 1, 2;, assuming the clock lane is on hardware lane 0.
+ This property is valid for serial busses only (e.g. MIPI CSI-2).
+- clock-lanes: an array of physical clock lane indexes. Position of an
+ entry determines the logical lane number, while the value of an entry
+ indicates physical lane, e.g. for a MIPI
Re: [PATCH RFC v4 01/14] [media] Add common video interfaces OF bindings documentation
Hi Sylwester,
On Thursday 24 January 2013 19:30:10 Sylwester Nawrocki wrote:
On 01/24/2013 11:16 AM, Laurent Pinchart wrote:
[...]
+Data interfaces on all video devices are described by their child 'port'
+nodes. Configuration of a port depends on other devices participating in
+the data transfer and is described by 'endpoint' subnodes.
+
+dev {
+ #address-cells = 1;
+ #size-cells = 0;
+ port@0 {
+ endpoint@0 { ... };
+ endpoint@1 { ... };
+ };
+ port@1 { ... };
+};
+
+If a port can be configured to work with more than one other device on
+the same bus, an 'endpoint' child node must be provided for each of
+them. If more than one port is present in a device node or there is more
+than one endpoint at a port, a common scheme, using '#address-cells',
+'#size-cells' and 'reg' properties is used.
Wouldn't this cause problems if the device has both video ports and a
child bus ? Using #address-cells and #size-cells for the video ports would
prevent the child bus from being handled in the usual way.
Indeed, it looks like a serious issue in these bindings.
A possible solution would be to number ports with a dash instead of a @,
as done in pinctrl for instance. We would then get
port-0 {
endpoint-0 { ... };
endpoint-1 { ... };
};
port-1 { ... };
Sounds like a good alternative, I can't think of any better solution at the
moment.
+Two 'endpoint' nodes are linked with each other through their
+'remote-endpoint' phandles. An endpoint subnode of a device contains
+all properties needed for configuration of this device for data exchange
+with the other device. In most cases properties at the peer 'endpoint'
+nodes will be identical, however they might need to be different when
+there is any signal modifications on the bus between two devices, e.g.
+there are logic signal inverters on the lines.
+
+Required properties
+---
+
+If there is more than one 'port' or more than one 'endpoint' node
following +properties are required in relevant parent node:
+
+- #address-cells : number of cells required to define port number,
should be 1.
+- #size-cells: should be zero.
I wonder if we should specify whether a port is a data sink or data
source. A source can be connected to multiple sinks at the same time, but
a sink can only be connected to a single source. If we want to perform
automatic sanity checks in the core knowing the direction might help.
Multiple sources can be linked to a single sink, but only one link can be
active at any time.
So I'm not sure if knowing if a DT port is a data source or data sink would
let us to validate device tree structure statically in general.
Such source/sink property could be useful later at runtime, when data
pipeline is set up for streaming.
Yes, I was mostly thinking about runtime.
How do you think this could be represented ? By just having boolean
properties like: 'source' and 'sink' in the port nodes ? Or perhaps in the
endpoint nodes, since some devices might be bidirectional ? I don't recall
any at the moment though.
Source and sink properties would do. We could also use a direction property
that could take sink, source and bidirectional values, but that might be more
complex.
I don't think we will have bidirectional link (as that would most probably
involve a very different kind of bus, and thus new bindings).
+Optional endpoint properties
+
+
+- remote-endpoint: phandle to an 'endpoint' subnode of the other device
+ node.
+- slave-mode: a boolean property, run the link in slave mode.
+ Default is master mode.
What are master and slave modes ? It might be worth it describing them.
This was originally proposed by Guennadi, I think he knows exactly what's
the meaning of this property. I'll dig into relevant documentation to
find out and provide more detailed description.
Thank you.
+- bus-width: number of data lines, valid for parallel busses.
+- data-shift: on parallel data busses, if bus-width is used to specify
+ the number of data lines, data-shift can be used to specify which data
+ lines are used, e.g. bus-width=10; data-shift=2; means, that
+ lines 9:2 are used.
+- hsync-active: active state of HSYNC signal, 0/1 for LOW/HIGH
+ respectively.
+- vsync-active: active state of VSYNC signal, 0/1 for LOW/HIGH
+ respectively. Note, that if HSYNC and VSYNC polarities are not
+ specified, embedded synchronization may be required, where supported.
+- data-active: similar to HSYNC and VSYNC, specifies data line polarity.
+- field-even-active: field signal level during the even field data
+ transmission.
+- pclk-sample: sample data on rising (1) or falling (0) edge of the
+ pixel clock signal.
+- data-lanes: an array of physical data lane indexes. Position of an
+ entry
[PATCH RFC v4 01/14] [media] Add common video interfaces OF bindings documentation
From: Guennadi Liakhovetski g.liakhovet...@gmx.de
This patch adds a document describing common OF bindings for video
capture, output and video processing devices. It is curently mainly
focused on video capture devices, with data busses defined by
standards like ITU-R BT.656 or MIPI-CSI2.
It also documents a method of describing data links between devices.
Signed-off-by: Guennadi Liakhovetski g.liakhovet...@gmx.de
Signed-off-by: Sylwester Nawrocki s.nawro...@samsung.com
Reviewed-by: Stephen Warren swar...@nvidia.com
Acked-by: Rob Herring rob.herr...@calxeda.com
---
Changes since v3:
- improved clock-lanes property description,
- grammar corrections of the example dts snippet description.
---
.../devicetree/bindings/media/video-interfaces.txt | 204
1 file changed, 204 insertions(+)
create mode 100644 Documentation/devicetree/bindings/media/video-interfaces.txt
diff --git a/Documentation/devicetree/bindings/media/video-interfaces.txt
b/Documentation/devicetree/bindings/media/video-interfaces.txt
new file mode 100644
index 000..0da126f
--- /dev/null
+++ b/Documentation/devicetree/bindings/media/video-interfaces.txt
@@ -0,0 +1,204 @@
+Common bindings for video data receiver and transmitter interfaces
+
+General concept
+---
+
+Video data pipelines usually consist of external devices, e.g. camera sensors,
+controlled over an I2C, SPI or UART bus, and SoC internal IP blocks, including
+video DMA engines and video data processors.
+
+SoC internal blocks are described by DT nodes, placed similarly to other SoC
+blocks. External devices are represented as child nodes of their respective
+bus controller nodes, e.g. I2C.
+
+Data interfaces on all video devices are described by their child 'port' nodes.
+Configuration of a port depends on other devices participating in the data
+transfer and is described by 'endpoint' subnodes.
+
+dev {
+ #address-cells = 1;
+ #size-cells = 0;
+ port@0 {
+ endpoint@0 { ... };
+ endpoint@1 { ... };
+ };
+ port@1 { ... };
+};
+
+If a port can be configured to work with more than one other device on the same
+bus, an 'endpoint' child node must be provided for each of them. If more than
+one port is present in a device node or there is more than one endpoint at a
+port, a common scheme, using '#address-cells', '#size-cells' and 'reg'
properties
+is used.
+
+Two 'endpoint' nodes are linked with each other through their 'remote-endpoint'
+phandles. An endpoint subnode of a device contains all properties needed for
+configuration of this device for data exchange with the other device. In most
+cases properties at the peer 'endpoint' nodes will be identical, however
+they might need to be different when there is any signal modifications on the
+bus between two devices, e.g. there are logic signal inverters on the lines.
+
+Required properties
+---
+
+If there is more than one 'port' or more than one 'endpoint' node following
+properties are required in relevant parent node:
+
+- #address-cells : number of cells required to define port number, should be 1.
+- #size-cells: should be zero.
+
+Optional endpoint properties
+
+
+- remote-endpoint: phandle to an 'endpoint' subnode of the other device node.
+- slave-mode: a boolean property, run the link in slave mode. Default is master
+ mode.
+- bus-width: number of data lines, valid for parallel busses.
+- data-shift: on parallel data busses, if bus-width is used to specify the
+ number of data lines, data-shift can be used to specify which data lines are
+ used, e.g. bus-width=10; data-shift=2; means, that lines 9:2 are used.
+- hsync-active: active state of HSYNC signal, 0/1 for LOW/HIGH respectively.
+- vsync-active: active state of VSYNC signal, 0/1 for LOW/HIGH respectively.
+ Note, that if HSYNC and VSYNC polarities are not specified, embedded
+ synchronization may be required, where supported.
+- data-active: similar to HSYNC and VSYNC, specifies data line polarity.
+- field-even-active: field signal level during the even field data
transmission.
+- pclk-sample: sample data on rising (1) or falling (0) edge of the pixel clock
+ signal.
+- data-lanes: an array of physical data lane indexes. Position of an entry
+ determines the logical lane number, while the value of an entry indicates
+ physical lane, e.g. for 2-lane MIPI CSI-2 bus we could have
+ data-lanes = 1, 2;, assuming the clock lane is on hardware lane 0.
+ This property is valid for serial busses only (e.g. MIPI CSI-2).
+- clock-lanes: an array of physical clock lane indexes. Position of an entry
+ determines the logical lane number, while the value of an entry indicates
+ physical lane, e.g. for a MIPI CSI-2 bus we could have clock-lanes = 0;,
+ which places the clock lane on hardware lane 0. This property is valid for
+ serial busses only (e.g. MIPI CSI-2). Note that for the MIPI CSI-2 bus this
+ array
